We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

stat.ME
< prev | next >
new | recent | 1702

Change to browse by:

stat

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo

Statistics > Methodology

Title: Test statistics and p-values

Authors: Yuri Gurevich, Vladimir Vovk
(Submitted on 8 Feb 2017 (v1), last revised 27 May 2017 (this version, v2))
Abstract: We point out that the traditional notion of test statistic is too narrow, and we propose a natural generalization that is arguably maximal. The study is restricted to simple statistical hypotheses.
Comments: 22 pages
Subjects: Methodology (stat.ME)
MSC classes: 62G10, 62F03, 62A01, 06A05
Journal reference: Proceedings of Machine Learning Research 105:89-104 (2019)
Cite as: arXiv:1702.02590 [stat.ME]
  (or arXiv:1702.02590v2 [stat.ME] for this version)

Submission history

From: Vladimir Vovk [view email]
[v1] Wed, 8 Feb 2017 19:27:44 GMT (29kb)
[v2] Sat, 27 May 2017 13:15:56 GMT (21kb)
Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)

Link back to: arXiv, form interface, contact.